Benjamin Negrevergne

Constraint-Based Sequence Mining Using Constraint Programming

The goal of constraint-based sequence mining is to find sequences of symbols that are included in a large number of input sequences and that satisfy some constraints specified by the user. Many constraints have been proposed in the literature, but a general framework is still missing. We investigate the use of constraint programming as general framework for this task.

Discovering closed frequent itemsets on multicore: Parallelizing computations and optimizing memory accesses

The problem of closed frequent itemset discovery is a fundamental problem of data mining, having applications in numerous domains. It is thus very important to have efficient parallel algorithms to solve this problem, capable of efficiently harnessing the power of multicore processors that exists in our computers (notebooks as well as desktops). In this paper we present PLCMQS, a parallel algorithm based on the LCM algorithm, recognized as the most efficient algorithm for sequential discovery of closed frequent itemsets. We also present a simple yet powerfull parallelism interface based on the concept of Tuple Space, which allows an efficient dynamic sharing of the work. Thanks to a detailed experimental study, we show that PLCMQS is efficient on both on sparse and dense databases.

Dominance Programming for Itemset Mining

Finding small sets of interesting patterns is an important challenge in pattern mining. In this paper, we argue that several well-known approaches that address this challenge are based on performing pair wise comparisons between patterns. Examples include finding closed patterns, free patterns, relevant subgroups and skyline patterns. Although progress has been made on each of these individual problems, a generic approach for solving these problems (and more) is still lacking. This paper tackles this challenge. It proposes a novel, generic approach for handling pattern mining problems that involve pair wise comparisons between patterns. Our key contributions are the following. First, we propose a novel algebra for programming pattern mining problems. This algebra extends relational algebras in a novel way towards pattern mining. It allows for the generic combination of constraints on individual patterns with dominance relations between patterns. Second, we introduce a modified generic constraint satisfaction system to evaluate these algebraic expressions. Experiments show that this generic approach can indeed effectively identify patterns expressed in the algebra.

Efficient Parallel Mining of Gradual Patterns on Multicore Processors

Mining gradual patterns plays a crucial role in many real world applications where huge volumes of complex numerical data must be handled, e.g., biological databases, survey databases, data streams or sensor readings. Gradual patterns highlight complex order correlations of the form “The more/less X, the more/less Y”. Only recently algorithms have appeared to mine efficiently gradual rules. However, due to the complexity of mining gradual rules, these algorithms cannot yet scale on huge real world datasets. In this paper, we thus propose to exploit parallelism in order to enhance the performances of the fastest existing one (GRITE) on multicore processors. Through a detailed experimental study, we show that our parallel algorithm scales very well with the number of cores available.

PGLCM: efficient parallel mining of closed frequent gradual itemsets

Numerical data (e.g., DNA micro-array data, sensor data) pose a challenging problem to existing frequent pattern mining methods which hardly handle them. In this framework, gradual patterns have been recently proposed to extract covariations of attributes, such as: “When X increases, Y decreases”. There exist some algorithms for mining frequent gradual patterns, but they cannot scale to real-world databases. We present in this paper GLCM, the first algorithm for mining closed frequent gradual patterns, which proposes strong complexity guarantees: the mining time is linear with the number of closed frequent gradual itemsets. Our experimental study shows that GLCM is two orders of magnitude faster than the state of the art, with a constant low memory usage. We also present PGLCM, a parallelization of GLCM capable of exploiting multicore processors, with good scale-up properties on complex datasets. These algorithms are the first algorithms capable of mining large real world datasets to discover gradual patterns.

PGP-mc: towards a multicore parallel approach for mining gradual patterns

Gradual patterns highlight complex order correlations of the form “The more/less X, the more/less Y”. Only recently algorithms have appeared to mine efficiently gradual rules. However, due to the complexity of mining gradual rules, these algorithms cannot yet scale on huge real world datasets. In this paper, we propose to exploit parallelism in order to enhance the performances of the fastest existing one (GRITE). Through a detailed experimental study, we show that our parallel algorithm scales very well with the number of cores available.

Modeling in MiningZinc

MiningZinc offers a framework for modeling and solving constraint-based mining problems. The language used is MiniZinc, a high-level declarative language for modeling combinatorial (optimisation) problems. This language is augmented with a library of functions and predicates that help modeling data mining problems and facilities for interfacing with databases. We show how MiningZinc can be used to model constraint-based itemset mining problems, for which it was originally designed, as well as sequence mining, Bayesian pattern mining, linear regression, clustering data factorization and ranked tiling. The underlying framework can use any existing MiniZinc solver. We also showcase how the framework and modeling capabilities can be integrated into an imperative language, for example as part of a greedy algorithm.